Aplicação de técnicas de machine learning na otimização da gestão hospitalar

This study aims to apply the K-Means clustering algorithm to identify disease patterns and support decision-making at Catapa Hospital. In this context, the integration of emerging technologies, particularly Artificial Intelligence and Machine Learning becomes essential for improving clinical data analysis and optimizing health resource administration. The methodology involved the collection and processing of 4,050 clinical records, covering variables such as date of attendance, neighborhood, gender, age, age group, type of disease, and neighborhood classification. After data preprocessing, the K-Means algorithm enabled the formation of five clusters composed of patients with similar characteristics. The analysis indicated that the Catapa neighborhood (classified as peri-urban) accounts for approximately 58% of malaria and fever cases. The youth age group represents 38.4% of the records, followed by children (32.3%), adolescents (14.9%), adults (11.8%), and the elderly (2.6%), with a slight predominance of the male gender (50.1%). Catapa alone accounts for 37% of recorded cases; Mbemba Ngango for 15%; Kindenuku for 10%; Dunga for 5%; Papelão for 3%; and the remaining 34 neighborhood